search

SciTools / iris-esmf-regrid

A collection of structured and unstructured ESMF regridding schemes for Iris.
https://iris-esmf-regrid.readthedocs.io/en/latest
BSD 3-Clause "New" or "Revised" License
19 stars 17 forks source link

Is the fix from PR 353 compatible with esmpy<8.6.0 as well? #415

Closed valeriupredoi closed 1 day ago

valeriupredoi commented 3 days ago

Hey folks, we are hitting some environment solving issues, and would like to use esmpy/esmf<8.6.0 (specifically, 8.4.2) but we are a bit scared that the fix from https://github.com/SciTools-incubator/iris-esmf-regrid/pull/353 will not work for those esmf/esmpy's - could one of you folks confirm we are good to use esmf/esmpy==8.4.2 and still benefit from the fix, please? Many thanks in advance! Cheers :beer:

stephenworsley commented 1 day ago

@valeriupredoi I believe this fix should be backwards compatible. ESMF became more strict about the format it expects indices to be given in (UGRID allows both 0 indexing and 1 indexing), this fix changes the format to be compatible with ESMF>8.6 while remaining otherwise equivalent. The only thing this fix changes is that the indexing format on the original cubes used passed down "as is" to ESMF, but I doubt that was ever relevant.

valeriupredoi commented 1 day ago

brilliant, thank you heaps @stephenworsley :beer: Phew! @schlunma we can then still be fine environment-wise, bud